A powerful integration for performing AI-driven web searches and structured content extraction via the inference.sh CLI.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install web-search-2
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install web-search-2 using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Web Search and Content Extraction?
The Web Search and Content Extraction skill enables developers and AI agents to access the live web through premium search engines like Tavily and Exa. By utilizing this skill within the Openclaw Skills ecosystem, you can bypass the knowledge cutoff of standard LLMs, allowing your applications to perform real-time research, verify facts, and scrape high-quality content for data processing pipelines.
This skill provides a unified interface for both search-to-answer workflows and raw content extraction. Whether you need a direct factual answer to a specific question or want to ingest multiple articles into a RAG system, these Openclaw Skills offer the necessary tools to build sophisticated, internet-aware agents.
Web Search and Content Extraction Use Cases
- Conduct deep research and gather the latest information on rapidly evolving topics.
- Power Retrieval-Augmented Generation (RAG) pipelines with fresh, relevant context from the web.
- Verify claims and perform automated fact-checking by fetching data from multiple sources.
- Aggregate and analyze content from specific URLs for market intelligence or academic study.
- Equip autonomous agents with the ability to browse the internet and perform multi-step research tasks.
How Web Search and Content Extraction Works
- Install the inference.sh CLI on your local machine or server environment.
- Authenticate the CLI using the login command to access the search provider APIs.
- Trigger a search or extraction task using either the Tavily or Exa applications via the infsh command.
- Receive structured JSON output containing search results, AI-generated summaries, or extracted page content.
- Optionally pipe the output into an LLM skill to synthesize the information or generate a final report.
Web Search and Content Extraction Setup
To get started with these Openclaw Skills, install the CLI and authenticate your session:
curl -fsSL https://cli.inference.sh | sh && infsh login
Run your first search using the Tavily Search Assistant:
infsh app run tavily/search-assistant --input '{"query": "latest AI developments 2024"}'
Web Search and Content Extraction Data Schema & Taxonomy
The skill utilizes a standardized input/output format for different search applications. Below is the mapping of common apps available in this skill set:
| App ID | Input Schema | Primary Output |
|---|---|---|
tavily/search-assistant |
{"query": "string"} |
AI-generated answer with sources and images |
tavily/extract |
{"urls": ["string"]} |
Clean text content and media from specified URLs |
exa/search |
{"query": "string"} |
Neural-ranked links with context snippets |
exa/answer |
{"question": "string"} |
Direct factual answers based on web data |
exa/extract |
{"url": "string"} |
Analyzed and parsed web page content |
Web Search and Content Extraction Advanced Features
- Multi-Provider Flexibility: Choose between Tavily's AI-optimized summaries or Exa's neural search for high-relevance link discovery.
- Seamless RAG Integration: Designed to work alongside LLM skills in Openclaw Skills to create end-to-end research workflows.
- Parallel Extraction: Extract content from multiple URLs in a single command using the Tavily Extract tool.
- Structured Data Output: All results are returned as parseable JSON, making it easy to integrate search data into automated scripts or databases.
- Agent-Ready Tools: Easily register these tools within agentic frameworks to give your AI the ability to use a search engine effectively.
SKILL.md
Loading
Related Openclaw Skills
Video Prompting Guide
A comprehensive framework for engineering high-quality prompts to generate cinematic AI videos using modern models.
Video Ad Specs
A professional toolkit for generating platform-optimized video ads with exact technical specifications for TikTok, Instagram, YouTube, and LinkedIn.
Tool UI Components
A professional collection of React and Next.js components designed to visualize the complete lifecycle of AI agent tool executions.
Text-to-Speech
A comprehensive text-to-speech skill for AI agents to generate natural, expressive, and multi-speaker audio via the inference.sh platform.
Widget Renderer
A declarative UI engine that renders complex React components from simple JSON structures provided by AI agents.
YouTube Thumbnail Design
Automate the creation of high-CTR YouTube thumbnails using AI image generation and data-driven design psychology.
METADATA
Featured*
